Examining robustness of bipedal robot walking with artificial sensitive skin

由于浮基系统的自然不稳定动力学，仿人机器人的两足平衡和运动是一项具有挑战性的任务。该项目名为“测试人工敏感皮肤行走的两足机器人的健壮性”，旨在使仿人机器人能够使用人工敏感皮肤在不平坦和狭窄的地形上行走。这将得到基于ROS的软件控制框架的开发的支持。人的脚底覆盖着一层耐用的无毛皮肤，除了保护脚免受粗糙表面的影响外，还提供了关于地形的丰富感官信息，如纹理、硬度、温度和立足点上的压力分布。今天，大多数类人机器人只配备了脚踝关节扭矩-力传感器，它们在运动时不能考虑任何类型的表面接触(任何类型的接触几何)。如果仿人机器人要更多地应用于为人类创造的环境中，它们需要能够在各种地面和表面上行走、平衡。在认知系统研究所过去几年开发的机器人皮肤的帮助下，一种在脚底安装了皮肤传感器的仿人机器人已经被开发出来。人工触觉传感器可以提供物理相互作用力的空间信息。这些信息可能有助于提高机器人系统和控制方法的能力。本项目的重点是研究足底触觉对持续行走控制的益处。在两个不同的步行控制器中，不同的机器人皮肤感知模式将被测试用于在不平坦的地形上行走，而不需要进行重大修改。我们将研究如何使用皮肤信息来在线调整行走动作，以踏上意想不到的障碍，特别是在小尺寸/狭窄立足点的情况下。将对足底接近感应器用于先发制人的足部顺应性进行调查，这可能提供一种感测方式，使脚能够在脚部着陆之前调整其方向，使用接近地面的情况作为控制的参考。该框架将在两个不同的全尺寸人形机器人上进行测试，这些机器人运行在两个不同的行走控制器上。

Humanoid robot's biped balance and locomotion is a challenging task due to the naturally unstable dynamics of floating base systems. The project “Examining robustness of bipedal robot walking with artificial sensitive skin” aims at enabling humanoid robots to walk over uneven and narrow terrains more robustly with artificial sensitive skin. This will be supported by the development of a ROS-based software control framework. The human foot sole is covered with an endurable layer of glabrous skin that apart from protecting the foot from rough surfaces, provides rich sensorial information about the terrain such as texture, hardness, temperature, and pressure distribution on the foothold. Today, most humanoids robots are only equipped with ankle joint torque-force sensors, they cannot account for any kind of surface contact (any kind of contact geometry) when making movements. If humanoid robots are to be used more in environments made for human beings, they need to be able to walk, balance on various kinds of ground and surfaces.With the help of the robotic skin developed by the Institute for Cognitive Systems over the past years, a humanoid robot equipped with skin sensors on the sole of its feet has been developed. The artificial tactile sensors can provide spatial information on the physical interaction forces. Such information can be useful to improve the capabilities of robotic systems and control methods. This project focuses on the investigation of the benefits that plantar tactile sensation provides for continuous walking control. Different sensing modalities of robot skin will be examined for walking on uneven terrain in two different walking controllers without significant modifications will be the target. We will investigate how skin information can be used to adapt online the walking motions for stepping on unexpected obstacles, especially in the case of small-size/narrow footholds. The use of plantar proximity sensing for pre-emptive foot compliance will be investigated, which could provide a sensing modality that may enable the foot to adapt its orientation before foot landing using the proximity-to-ground as a reference for control. The framework will be tested in two different full-size humanoid robots running on two different walking controllers.